This paper presents a WordNet-based automatic approach for doing “easy” inferences. We build an automatic system which extracts the pairs of the SICK corpus whose sentences only differ by one or two words and then identifies which inference relation (i.e. entailment, contradiction, neutrality) exists between these words, based onWordNet relations. Since the sentences of those pairs only differ by these words of the comparison, the inference relation found between the words is taken to apply on the whole sentences of the pair. For some cases not dealt by WordNet we use our own heuristics to label the inference type. With our approach we accomplish three goals: a) we manage to correct the annotations of a part of the SICK corpus and provide the corrected corpus, b) we evaluate the coverage and relation-completeness of WordNet and provide taxonomies of its strengths and weaknesses and c) we observe that “easy” inferences are a suitable evaluation technique for lexical resources and suggest that more such methods are used in the task. The outcome of our work can eventually help improve the SICK corpus and the WordNet resource and it also introduces a new way for dealing with lexical resources evaluation tasks.
@InProceedings{KALOULI18.18, author = {Aikaterini-Lida Kalouli ,Livy Real and Valeria DePaiva}, title = {WordNet for “Easy” Textual Inferences}, booktitle = {Proceedings of the Eleventh International Conference on Language Resources and Evaluation (LREC 2018)}, year = {2018}, month = {may}, date = {7-12}, location = {Miyazaki, Japan}, editor = {}, publisher = {European Language Resources Association (ELRA)}, address = {Paris, France}, isbn = {979-10-95546-28-3}, language = {english} }
Summary of the paper